autopilot_skill
- TypeScript
9.3k
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill yeachan-heo/oh-my-claudecode --skill autopilot- SKILL.md8.5 KB
Overview
This skill provides full autonomous execution from a short idea to verified, working code. It orchestrates requirements, technical design, parallel implementation, QA cycles, and multi-perspective validation so you get a completed project without managing individual steps. It is teams-first and built to run unattended until completion or explicit cancellation.
How this skill works
You give a 2–3 line product description and the skill expands it into a detailed spec, generates an implementation plan, runs parallel executors to build code, and performs iterative QA until tests and builds pass. A final validation phase runs independent architecture, security, and code reviews; all reviewers must approve before cleanup. Configurable limits guard repeated failures and let you pause, cancel, or resume progress.
When to use it
- You want end-to-end autonomous delivery from idea to working code.
- You tell the system to “autopilot”, “build me”, “create me”, or a similar hands-off trigger.
- The task requires multiple phases: planning, parallel implementation, testing, and validation.
- You prefer the system to run to completion without coordinating each step.
- You can accept automated decision-making and occasional prompts for vague or missing requirements.
Best practices
- Provide a concise but specific product description: domain, core features, and tech constraints (e.g., TypeScript, DB).
- Prefer feature-rich tasks that need coordinated phases rather than single-line fixes or brainstorming.
- Allow the system to run; avoid interrupting unless you need to cancel or provide clarifications.
- Use configuration options to adjust max QA cycles, validation rounds, or to pause between phases.
- If QA errors repeat or validation stalls, inspect logs and provide clearer requirements or resume after edits.
Example use cases
- Autonomously build a TypeScript REST API for a bookstore with CRUD and tests from a short prompt.
- Create a CLI habit tracker with streak counting, implemented and validated end-to-end.
- Take a clear feature request and deliver a finished module with tests, linting, and multi-perspective reviews.
- Run unattended development for a small product proof-of-concept that must include security and code-quality checks.
FAQ
QA cycles repeat up to a configured limit. If the same error persists across 3 cycles, the run stops and reports the fundamental issue for human intervention.
Can I cancel or resume a run?
Yes. Use the cancel command to stop and preserve progress, and restart the autopilot command to resume from the last saved state.
Will it block if external validation tools are unavailable?
No. The system tries to use external Codex/MCP tools but proceeds without them if they are unavailable; it never blocks indefinitely.